Shane Tan.

Home / Work

Work — in detail

Three projects in depth and two hackathon results. Every entry separates what I built or contributed from what the team built, and states its status beside it.

Featured
3 projects · 2026
Additional
2 hackathon results · 2026
Kinds
Team prototypes, a team proof of concept and hackathon projects
Focus
Multi-agent reliability, human review, evaluation
  • Built by me or my contribution
  • Team system
Artwork
Artwork Abstract: one request, a gate, the agents beyond it. Not the system architecture.

No. 01Multi-agent AI · 2026

finAI — the coordination layer

How a rule-first Clarifier and a deterministic Orchestrator keep a team's specialist agents inside policy.

Read the case study
My role
Clarifier and Orchestrator coordination layer
Context
SMU IS4000 team project · wealth-advisory use case
Status
Team prototype · specialist providers in test mode

Built by me

  • Rule-first Clarifier
  • Deterministic plan admission and the Orchestrator runtime
  • Single-use approval cards

Team systemBuilt by teammates

  • Specialist agents
  • Application graph and database checkpointing
Artwork
Artwork Abstract: sources converging at a checkpoint that can pause a case. Not the system architecture.

No. 02Corporate onboarding · 2026

Corporate Onboarding

A recoverable case workflow: when evidence conflicts, the case pauses, a person reviews it, and it resumes from where it stopped.

My role
Co-developer in a six-person team
Context
Industry-sponsored final year project
Status
Team proof of concept · mock integrations · not production

My contribution

  • Co-developed the team's proof of concept
  • Contributed to the review-and-recovery workflows

Team systemSix-person team

  • Registry lookup (mock/sandbox)
  • OCR-assisted extraction
  • Reconciliation and ownership analysis
Read the summary
Artwork
Artwork Abstract, not real footage or a system output.

No. 03Computer-vision review · 2026

SafeTrace

A local-first prototype that reviews visual safety evidence on the user's own computer and links every finding back to its source frames.

My role
AI Solutions Intern · integration and connector work
Context
Internship prototype in a team repository
Status
Local-first prototype · not production

My contribution

  • React/FastAPI integration
  • Local workflow hardening
  • Evidence-review experience
  • Hosted-to-local connector

Team systemBuilt by the team

  • Frame retrieval
  • Vision models
  • Rule layer
Read the case study

Additional work

Hackathon projects · team results

04

LoreDrop

2nd place, Advanced category · CloudHacks 2026 · four-person team

A hackathon learning app that turns study material into an audio-led learning flow, with summaries, attention checks, a quiz, a recap and grounded Q&A.

My contribution

  • The grounded Q&A flow, with its retrieval, scope resolution, provenance and tests

Team system

  • Content generation, attention checks and audio
05

Team RedMana

3rd place, Learners category · SHINE 2026

Team member.